Study of mechanical damage to tomato seedling stems to ensure the quality of grafting

Abstract

Tomato seedling stems can suffer mechanical damage during automatic grafting, reducing the quality of grafting. This study examined two sources of damage: compression damage to scion stems and bending damage to rootstock stems. Two damage tests were conducted. The test factors were determined based on the finite element simulation method. The four test indicators were established through a stress analysis method. Based on the damage test, for seedlings aged 33–35 days, when the diameter compression ratio was less than 40% or when the deviation distance of the rootstock emergence point in the longitudinal and transverse directions on the substrate bowl surface was less than 10 mm, the equivalent bending and buckling forces significantly exceeded the weight (0.0228 N) imposed by the graft clip and scion. Consequently, the deformation distance between the mass centre of the grafted seedling and the centre of the substrate bowl decreased to 0 mm, and the grafting success rate was 100%, with upright grafted seedlings and high–quality graft formation. This study provides a foundation for optimising clamping and alignment mechanisms in grafting machines.

Keywords:
tomato seedlings; grafting; compression damage; bending damage; finite element

Nomenclature

Symbols Symbols h Stem height, mm Δx The deviation distance in the transverse direction (x–axis), mm d Hypocotyl diameter of the rootstock, mm Δz The deviation distance in the longitudinal direction (z–axis), mm E Elastic modulus of the rootstock hypocotyl, MPa L Deformation distance, the horizontal distance between the mass centre of the grafted seedling and the centre of the substrate bowl, mm F Bending force in the three point bending test, N G Weight of the grafted seedling, N Bending deflection at the loading point, mm F A Bending force at the elastic limit, N l Span length between supports in the three–point bending test, mm w A Deflection at the elastic limit, mm I Area moment of inertia of the hypocotyl cross–section, mm4 F B Bending force at the bending strength, N k slope of the elastic region in the force–displacement curve, N/mm w B Deflection at the bending strength, mm H y Vertical distance from the alignment bar to the substrate bowl surface, mm M Damage force, the maximum von Mises stress in the rootstock alignment simulation test, N H y2 Vertical distance from the alignment bar to the centring bar, mm GSR Grafting success rate, % T y Vertical thickness of the alignment bar, mm RR Rate of reduction of the equivalent bending or buckling force, % T y2 Vertical thickness of the centring bar, mm

Symbols		Symbols
h	Stem height, mm	Δx	The deviation distance in the transverse direction (x–axis), mm
d	Hypocotyl diameter of the rootstock, mm	Δz	The deviation distance in the longitudinal direction (z–axis), mm
E	Elastic modulus of the rootstock hypocotyl, MPa	L	Deformation distance, the horizontal distance between the mass centre of the grafted seedling and the centre of the substrate bowl, mm
F	Bending force in the three point bending test, N	G	Weight of the grafted seedling, N
	Bending deflection at the loading point, mm	F A	Bending force at the elastic limit, N
l	Span length between supports in the three–point bending test, mm	w A	Deflection at the elastic limit, mm
I	Area moment of inertia of the hypocotyl cross–section, mm4	F B	Bending force at the bending strength, N
k	slope of the elastic region in the force–displacement curve, N/mm	w B	Deflection at the bending strength, mm
H y	Vertical distance from the alignment bar to the substrate bowl surface, mm	M	Damage force, the maximum von Mises stress in the rootstock alignment simulation test, N
H y2	Vertical distance from the alignment bar to the centring bar, mm	GSR	Grafting success rate, %
T y	Vertical thickness of the alignment bar, mm	RR	Rate of reduction of the equivalent bending or buckling force, %
T y2	Vertical thickness of the centring bar, mm

Introduction

Tomatoes have significant nutritional and economic importance throughout the world (Burusa et al., 2024; Gao et al., 2024; Wang et al., 2024). In 2023, global tomato production surpassed 192 million tonnes, yielding an average of 35 tonnes of fruit per hectare, and this yield is expected to increase further (Fao, 2025). To meet this rising production demand, tomato growers aim to scale up production (Asadbeigi et al., 2023; Burusa et al., 2024; Medland, 2021). Besides, grafting technology is widely applied in tomato production to overcome continuous cropping obstacles and to control soil–borne diseases, thus increasing the yield (Chen et al., 2020; Jiang et al., 2022; Yan et al., 2022). However, grafting is labour intensive and involves tedious, repetitive actions (Comba et al., 2016). With the development of large–scale and standardised production of tomato seedlings, there is an urgent need for tomato grafting machines, which can significantly reduce labour intensity, expand production capacities, and improve product uniformity (Liang et al., 2023). One important factor limiting the practical application of grafting machines is the lack of uniform seedlings that are suitable for these machines. Tomato seedlings with different diameters or random emergence points in the plug tray can be easily damaged during automatic grafting, which directly reduces the quality of grafting. Therefore, it is necessary to study the impact of the mechanical damage to tomato seedlings during automatic grafting to improve the application of grafting machines.

Previous research on mechanical damage in tomato seedlings has primarily focused on identifying and detecting its occurrence in the leaves and roots (Cheng et al., 2017; Liu et al., 2023a; Nifakos, 2022; Sun et al., 2016). However, mechanical damage to the stem is a critical factor that affects seedling quality in the context of automatic grafting. Such damage frequently results from compressive and bending stresses imposed by clamping and holding units. Yet, the existing literature on the relationship between stem damage and these stresses is limited. Several studies have investigated stem bending caused by natural wind disturbances, focusing on its long–term effects on plant morphology and physiology (e.g., Liu et al., 2023b; Yang et al., 2024). Meanwhile, another stream of research has addressed stem damage during transplanting operations, with the findings being largely applied to optimise the gripper design and to minimise harm (Jin et al., 2018). This area shares a somewhat close objective with automatic grafting: to design machinery that handles seedlings gently. However, a critical gap remains: although the objectives are similar, seedlings at the grafting stage are younger and consequently more prone to damage than those at the transplanting stage (Gallegos Cedillo et al., 2024; Kwack et al., 2014). Similarly to transplanting operations, studies on mechanical damage during grafting have mainly focused on machine design (Wu et al., 2021; Xie et al., 2022), while the impact of the physical attributes of seedlings on mechanical damage have been less explored.

Therefore, this study investigated the causes of mechanical damage to seedlings during grafting from the perspective of individual differences in seedling diameter and the emergence position of seedlings in the substrate trays. First, the key sources of damage arising from seedling morphological characteristics were identified. Then, the findings were used to optimise the design and operating parameters of the grafting machinery, ultimately ensuring high grafting quality while reducing mechanical damage. This research provides a critical foundation for the development of high–precision automatic grafting machines that exert minimal damage on tomato seedlings. Specifically, it makes two contributions: (1) this study investigates seedling mechanical damage from the perspective of their physical properties (e.g., variation in diameter and off–centre emergence) rather than from the perspective of the machinery’s design. (2) The criterion for mechanical damage is based on post–grafting upright seedlings rather than traditional severe damage (e.g., stem breakage).

Material and Methods

Physical parameters of tomato seedlings and the grafting clip

Red No. 2 tomato scions and T17–2 tomato rootstocks, provided by the Guangdong Provincial Improved Variety Introduction Service Corp., were used for the experiments. All seedlings were cultivated in standardised 72–cell trays (6 × 12 configuration), which were filled with a uniformly mixed substrate composed of peat, vermiculite, and perlite at a 7:2:1 ratio. Germinated seeds were sown into the trays on 16 August 2024 (scions) and 15 August 2024 (rootstocks), and then covered with substrate. The temperature of the greenhouse was maintained at 25–33°C during cultivation. Seedlings exhibiting stems with 3–6 compound leaves and a single apical bud were identified as meeting the mechanical grafting standards and utilised experimentally. A total of 20 tomato scions and 20 tomato rootstocks, each 28 days old, were selected randomly for the assessment of their physical parameters. Additionally, 33– and 38–day–old tomato seedlings were selected and measured for comparison. The parameters of the scion seedlings, including stem height, epicotyl length, epicotyl diameter, stem length above the cotyledon, and scion weight, were measured. The parameters of the rootstock seedlings, including stem height, hypocotyl length, and hypocotyl diameter, were also measured. Besides, the grafting clip weight was measured. The results are summarised in Table 1.

Table 1
Physical parameters of tomato seedlings and the grafting clip.

Scion						Grafting clip	Rootstock
Seedling age, days	Stem height h, mm	Epicotyl length, mm	Epicotyl diameter, mm	Stem length above cotyledon, mm	Scion weight, N	Clip weight, N	Stem height h, mm	Hypocotyl length, mm	Hypocotyl diameter d, mm
28	90.14±9.58	25.29±4.46	3.02±0.22	46.43±6.85	0.0109		84.43±9.93	36±7.57	2.75±0.09
33	104.86±6.89	28.43±3.55	3.62±0.19	61.43±5.47	0.0189	0.0033	73.14±3.89	35.57±3.91	2.97±0.25
38	102±8.23	21.71±4.86	3.79±0.17	55.57±2.99	0.0195		95±8.41	32.13±8.40	3.64±0.14

• Note: The data are presented as the mean ± standard deviation.

Structure and operational principle of the grafting machine

The HPGR high–speed grafting machine (Fig. 1) was used to investigate the interaction between the machine and tomato seedling stems. This machine is capable of grafting six seedlings simultaneously, achieving a productivity rate of 2,600 seedlings per hour with two operators responsible for feeding the seedlings. This machine comprises a seedling supply mechanism, an alignment mechanism, a cutting mechanism, a clip–feeding mechanism, a conveying mechanism, a replanting mechanism, and a frame (Fig. 2). The seedling supply mechanism includes a scion supply mechanism and a rootstock supply mechanism, where the former uses two semi–circular clamping hands to securely hold the scion stem and the latter positions the rootstock in a clamping hole mounted on a conveyor chain for transportation.

Figure 1
A diagram of the overall structure of the high–speed grafting machine.

Figure 2
The main operation process of the high–speed grafting machine.

The operational principle is as follows (Fig. 2): (i) the scion seedling without the hypocotyl is artificially sent to the scion supply mechanism, and the substrate bowl of the rootstock is clamped by a single manipulator into the clamping hole as the same time. (ii) The conveying mechanism transports the scion and rootstock to the designated position. (iii) The alignment mechanism precisely gathers the scion and rootstock in a fixed grafting position. (iv) The cutting mechanism cuts the lower end of the scion stem and the upper end of the rootstock stem into a 30° incision in the vertical direction. (v) The scion and rootstock are bonded by the grafting clip supplied by the clip–feeding mechanism. (vi) The conveying mechanism transfers the seedling after grafting to the replanting mechanism. (vii) The replanting mechanism replants the seedling after grafting into trays. (viii) Once all the grafted seedlings in one tray are prepared, they are manually transported to the healing room for recovery, completing the grafting operation.

Experimental design

Three–point bending tests

The mechanical parameters were derived from three–point bending tests on rootstock hypocotyls by an electric tensile testing machine manufactured by Dongguan Zhi Taking Precision Instrument Company. Based on the principles of engineering mechanics, the elastic modulus of the rootstock hypocotyl was determined using the following formula:

$E=\frac{F{l}^3}{48wI}=\frac{F}{w}\times \frac{l^3}{48I}$

Where:

E is the elastic modulus of the rootstock hypocotyl (MPa);

F is the bending force applied to the stem (N);

w is the bending deflection at the loading point (mm);

l is the span length between two support points in the three–point bending test (mm), and

I is the area moment of inertia of the hypocotyl cross–section (mm⁴).

The rootstock stem is considered to be a round solid rod, then:

$I=\frac{\pi d^4}{64}$

Where:

d represents the hypocotyl diameter of the rootstock (mm). E can then be expressed as follows:

$E=k\times \frac{4l^3}{3\pi d^4}$

Where:

k = F/w is defined as the slope of the elastic region in the force–displacement curve (N/mm).

During the test, the span length l was set to 20 mm. Ten rootstock hypocotyls were selected for the experiment, and their elastic regions were fitted by the least square fitting method.

Grafting test

Based on the operational principles of the grafting machine, it is evident that mechanical damage occurs during the clamping of the scion and the alignment and centring of the rootstock. To investigate the quantitative effects of such mechanical damage to the scion and rootstock stems on grafting quality, a preliminary grafting test was conducted. The experimental indicators are the grafting success rate and the grafting failure rate. The grafting success rate is defined as the proportion of the number of grafted seedlings without apparent morphological defects relative to the total number of seedlings. The sum of the grafting success rate and the grafting failure rate equals 1. This experiment was conducted using our high–speed grafting machine. A total of 360 scions and 360 rootstocks were selected for grafting. The grafting process was recorded using a real–time video, and the causes of grafting failure were subsequently analysed.

Rootstock alignment simulation test

During the grafting process, bending damage comes from the interaction between the rootstock stem and the alignment mechanism, which involves the alignment and centring process, as shown in Fig. 3. Specifically, the alignment bar initially corrects the rootstock that deviates from the grafting position in the x direction, causing bending damage in that direction. Subsequently, the alignment bar further corrects the rootstock in the z direction, leading to additional

Figure 3
The alignment and centring of the alignment mechanism.

bending damage in that direction. Once the positioning is complete, a 4 mm × 4 mm rectangular space is formed in the centre of the alignment bar. This space is larger than the maximum rootstock diameter of 3.64 mm, thereby preventing compression damage to the rootstock at this stage. Finally, the centring bar bends the upper portion of the rootstock to ensure positioning at x = 0, preparing it for subsequent cutting and clip–feeding operations. The factors that influence this process include the bending heights (H_y, H_y2), the bending thicknesses (T_y, T_y2) and the deviation distances (Δx, Δz). The bending height H_y is the vertical distance from the centre of the alignment bar to the substrate bowl surface. The bending height H_y2 refers to the vertical distance from the centre of the alignment bar to the centre of the centring bar. The bending thickness T_y is the vertical thickness of the alignment bar. The bending thickness T_y2 refers to the vertical thickness of the centring bar. The deviation distance Δx refers to the distance in the transverse direction (x–axis) and the deviation distance Δz is the distance in the longitudinal direction (z–axis) on the substrate bowl upper surface.

An alignment simulation test was carried out to identify the parameters that significantly influence rootstock stem bending damage. This test clearly visualised the relationship between rootstock stems and the alignment mechanism. According to the mechanism design requirements, the factors included H_y of 20–30 mm, H_y2 of 8–10 mm, and T_y and T_y2 of 2–6 mm. The results from three–point bending tests on the rootstock showed that under bending strength, the bending deflection w_B was less than 6 mm in most cases (Table 2). Accordingly, Δx and Δz varied from 2 to 6 mm. The seedling age was 33 days. Damage force M is defined as the maximum von Mises stress on the rootstock stem after alignment, was used as a direct measure of the degree of damage. The Plackett–Burman experimental design was applied, and a total of 12 trials were performed.

Table 2
Three–point bending test results.

No.	Span length l, mm	Hypocotyl diameter d, mm	Force FA, N	Displacement wA, mm	Force FB, N	Displacement wB, mm	Slope K, N/mm	Elastic modulus E, MPa
1	20	3.5	1.28	0.83	1.98	5.47	0.94	21.37
2	20	3.6	1.29	0.74	2.14	4.57	1.09	22.08
3	20	3.6	1.88	1.96	2.81	6.03	0.74	14.86
4	20	3.2	1.11	0.97	1.55	4.15	0.63	20.49
5	20	3.5	1.49	0.7	2.51	6.38	1.44	32.54
6	20	3.5	1.26	0.77	2.08	4.3	1.09	24.68
7	20	3.4	1.24	1.11	1.87	5.15	0.65	16.53
8	20	3.5	1.14	0.81	1.81	4.37	0.85	19.34
9	20	3.2	0.82	0.68	1.25	4.75	0.47	15.28
10	20	3.2	1.18	1	1.55	4.14	0.66	21.37

• Note: F_A and w_A represent the force and deflection at the elastic limit, while F_B and w_B denote the force and deflection at the bending strength.

Damage tests involving tomato seedling stems

Damage tests were conducted on tomato seedling stems to clarify the relationships between the operation parameters, seedling diameter, seedling age, and mechanical damage, and thereby optimise grafting quality.

Test indicators and scheme

Mechanical damage reduces the bearing capacity of the grafted seedling stem, resulting in bending deformation of the grafted seedling. The stress analysis of seedling stems is presented in Fig. 4. The distance between the centre of mass P of the grafted seedling and the centre O of the substrate bowl surface is denoted as L, defined as the deformation distance. This index provides a direct measure of the morphological characteristics of grafted seedlings. When there is no mechanical damage or only minor mechanical damage to the stem of the grafted seedlings, L is 0 mm. At this point, the grafted seedlings maintain an upright posture, resulting in a 100% grafting success rate, ensuring high–quality graft formation.

Figure 4
Stress analysis of grafted seedling deformation

When significant mechanical damage occurs, L exceeds 0 mm, and the grafting success rate is < 100%. For L > 0 mm, grafting failure can occur as follows: when P and O are misaligned, the grafted seedling stem will exhibit bending deformation due to its weight G; when P and O are aligned, the stem will experience buckling under the axial load of its weight G.

Mechanical analogies were employed to theoretically analyse these failure modes. For the bending case (Fig. 4a), the stem is represented by a cantilever beam. The bending moment M_bending induced by the weight is M_bending = G × L. For the buckling case (Fig. 4b), the stem can be modelled as a slender column with pinned ends. According to Euler’s buckling theory, the critical axial load P_buckling that causes elastic instability is given by P_buckling = (π2EI)/(Kh)2, where K is the effective length factor.

The core of the failure mechanism is that mechanical damage that reduces the stem’s intrinsic ability to withstand these critical loads. Therefore, the equivalent bending force (corresponding to the bending resistance capacity) and equivalent buckling force (corresponding to the critical buckling load P_buckling) are used as the test indicators. Grafting failure occurs when the stem’s load bearing capacity, quantified by these equivalent forces, is reduced due to damage to a level below the applied load from the plant’s weight under a specific posture. Besides, when the reduction in these forces is lower, there is less impact from mechanical damage, thereby ensuring better grafting quality. The rate of reduction is calculated as the ratio of the difference between the pre–damage and post–damage forces to the pre–damage force. Both forces and their rates of reduction serve as critical indicators for assessing the stem bearing capacity and grafting quality.

The damage test scheme is categorised into two types corresponding to the above two damage cases, as shown in Fig. 5. For the scion bending case, a tensile testing machine was utilised to simulate the clamping compression injury of the epicotyl. L was measured after the grafting the injured scion onto the rootstock. Subsequently, the epicotyl of the scion seedling was excised, and then the equivalent bending force and equivalent buckling force were measured. For the rootstock bending case, the portion above the cotyledon was removed. The hypocotyl, along with the substrate bowl, was damaged through precise alignment and centring. Subsequently, the injured rootstock and scion were grafted, and L was measured. Finally, the rootstock stem segment was partially excised from the substrate bowl, and the equivalent bending buckling forces of the hypocotyl were measured.

Figure 5
The damage test scheme for the tomato seedling stem

To characterise the bending and buckling deformation processes of damaged seedling stems, the equivalent bending and buckling forces were measured using a tensile testing machine. The analysis software included with the machine displayed the bending and buckling force data in real time; the data were subsequently exported for further processing. The sample length of the seedling stem was 40 mm, with the injury point located at the midpoint of the stem. During the damage tests, the clamping length of the fixed end of the stem segment was set to 5 mm. The equivalent bending force was defined as the peak force exerted when the free end of the stem segment was bent upward by a distance of 7 mm. The equivalent buckling force represented the maximum force applied when the free end of the stem segment was compressed downward by a distance of 10 mm during the buckling process.

Test factors and details

According to the damage test scheme, radial compression was employed to simulate clamping damage in scions. For rootstocks, stem bending and damage were induced using the alignment and centring method. The test factors and details are described below.

Compression damage test for scion

The diameter compression ratio, defined as the compression distance divided by the initial epicotyl diameter, was selected as the key test parameter. It was set at 0%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, and 60%, with 0% serving as the control group without any compression damage. The compressive strength of scion stems varies depending on the seedling age; therefore, the ages of 28, 33, and 38 days post–germination were selected for grafting, corresponding to an epicotyl diameter of 3.02, 3.62, and 3.79 mm, respectively. A standard indenter with an 8 mm diameter was used, and the compression speed was set at 10 mm/min. The compression duration was 6 s.

A total of 56 treatment groups were utilised in the experiment, with five replicates each. The stem deformation characteristics, compressive strain, compression force, L, grafting success rate, equivalent bending force, and equivalent buckling force were measured. Specifically, the deformation characteristics and compressive strain were measured in real time using a high–speed camera during the compression process. The compression force refers to the peak force experienced by the scion stems and was measured using the tensile testing machine.

Bending damage test for rootstock

To ensure adequate working space for the cutting and clip–feeding mechanisms, a minimum Hy of 20 mm was used in the bending damage test. Stem damage is most severe at this height. If grafting quality can be maintained at this height, then increasing Hy may further improve quality by reducing mechanical stress. Given that the maximum cell size in the 72–cell tray is 40 × 40 mm, the maximum of Δx and Δz was set to 20 mm. In the test, the Δx and Δz settings were (0, 0), (10, 10), and (20, 20) mm, with (0, 0) mm serving as the control group without bending damage. Seedlings of different ages exhibit varying resistance to bending damage, a factor that affects grafting quality. Three seedling ages were considered: 28, 33, and 38 days. The non–significant factors were set as follows: Ty and Ty2 were set to 2 mm, and Hy2 was set to 10 mm.

The experiment included 14 treatment groups, with five replicates each. L, the grafting success rate, equivalent bending force, and equivalent buckling force were measured in real time during the test.

Results and discussion

Mechanical parameters and rootstock stem finite model

The three–point bending test results for the 10 groups are presented in Table 2. The three–point bending force–displacement curve of the rootstock stem is presented in Fig. 6a, and the curve fitting for the elastic phase is illustrated in Fig. 6b. The average E for the 10 groups was 20.85 MPa. The Poisson ratio of the rootstock seedling was 0.3 (Liu et al., 2023).

Figure 6
The results of the tree–point bending test of the rootstock: (a) the fore–displacement curve; (b) elastic phase fitting; (c) an image of the physical test and a schematic of the simulation

An accurate finite element model of rootstock seedlings is essential to study mechanical damage to rootstocks. Thus, a three–point bending simulation model was established under identical test conditions (Fig. 6c). The peak value obtained from the three–point bending and simulation tests served as the reference. As illustrated in Fig. 6a, the relative error of the peak load between the simulation and physical tests was 0.6%. The results indicate that the simulation model accurately reflects the deformation behaviour of the rootstock.

Impact of mechanical grafting damage on the grafting success rate

The automatic grafting process produced four types of grafted seedlings: upright grafted seedlings without obvious morphological defects, grafted seedlings without a scion, grafted seedlings with a bent scion, and grafted seedlings with a bent rootstock. Grafted seedlings with scion or rootstock bending are prone to interference between plants or between plants and machinery during replanting, and these grafted seedlings are less favoured by the market. Therefore, grafted seedlings without a scion or with scion or rootstock bending are classified as grafting failures, while those without apparent morphological defects are classified as grafting successes. As presented in Table 3, the grafting success rate was 93% and the grafting failure rate was 7%. Among the failed grafts, absence of the scion accounted for 4%, scion bending for 2%, and rootstock bending for 1%.

Table 3
Grafting test results.

Test indicators	Grafting success rate, %	Grafting failure rate, %
Proportion, %	93	7
Damage	No damage	Other damage	Mechanical damage
Description of the damage	Grafted seedling without obvious morphological defects	Absence of the scion	Scion bending	Rootstock bending
Proportion of damage, %	0	4	2	1

Further analysis revealed that absence of the scion is primarily attributed to a significant diameter discrepancy between the rootstock and the scion. Specifically, when the diameter of the scion exceeds that of the rootstock, the grafting clip can only secure the larger–diameter scion stem during fixation. This creates a gap between the rootstock stem and the clip, resulting in insufficient clamping force. Consequently, both the scion and the grafting clip may detach, leading to grafted seedlings without scions. Conversely, when the rootstock diameter is larger than that of the scion, the clip similarly fails to secure both stems, resulting in the same detachment phenomenon.

Scion bending in grafted seedlings is attributed to mechanical damage during stem clamping. As illustrated in Table 1, there are significant variations in the epicotyl diameter of scions at different seedling ages. For scion seedlings aged 28–38 days, the corresponding epicotyl diameters ranged from 3.02 to 3.79 mm. The scion seedling supply mechanism employed a semi–circular clamping hand with a fixed 3 mm inner diameter to securely grasp the stems. However, this diameter was smaller than the actual scion epicotyl diameter (3.02–3.79 mm), resulting in unavoidable radial compression damage during clamping. This damage compromises the structural integrity of the scion stem, causing the stem to bend under the weight of the upper leaves.

The rootstock in grafted seedlings can be bent due to mechanical damage sustained by the rootstock stem during the alignment and centring process. Rootstock seedlings were supplied as seedlings grown in substrate bowls, retaining only the hypocotyl. As shown in Fig. 2, the emergence positions of rootstock seedlings within the substrate bowls are randomly distributed, with some deviating from the theoretical grafting position. During the alignment process, the mechanism shifted the rootstock stems to the centre of the substrate bowl for grafting, subjecting them to complex bending stresses that caused damage. This damage makes the rootstock stem unable to support the weight of the scion and graft clip, resulting in a grafted seedling with a bent rootstock.

The effects of different alignment parameters on the rootstocks

Tables 4 and 5 present the effects of alignment parameters on the rootstocks. As illustrated in Table 5, the model is statistically significant (P = 0.005), and the coefficient of determination (R²) is 0.9456, suggesting a good fit between the variables. Using multiple regression analysis, we derived the following regression equation:

Table 4
The design and results of the alignment simulation test.

	Bending height		Bending thickness		Deviation distance		Damage force
No.	A: Hy, mm	B: Hy2, mm	C: Ty, mm	D: Ty2, mm	E: Δx, mm	F: Δz, mm	M, N
1	30	10	2	6	6	6	0.8022
2	20	10	6	2	6	6	1.865
3	30	8	6	6	2	6	0.6385
4	20	10	2	6	6	2	1.440
5	20	8	6	2	6	6	1.871
6	20	8	2	6	2	6	1.008
7	30	8	2	2	6	2	0.6746
8	30	10	2	2	2	6	0.4624
9	30	10	6	2	2	2	0.2256
10	20	10	6	6	2	2	0.5641
11	30	8	6	6	6	2	0.7596
12	20	8	2	2	2	2	0.4035

Table 5
The results of variance analysis for the damage force.

Index	Source	Sum of squares	Degrees of freedom	F-value	P-value
Damage force M	Model	3.15	6	14.488	0.005*
	A: Hy	1.07	1	29.620	0.003*
	B: Hy2	1.401E-006	1	0.000	0.995
	C: Ty	0.107	1	2.953	0.146
	D: Ty2	0.007	1	0.193	0.679
	E: Δx	1.408	1	38.856	0.002**
	F: Δz	0.555	1	15.306	0.011*
	Residual	0.181	5
	Cor total	3.331	11
	R 2	0.9456
	Adjusted R2	0.8803

• Note: * P < 0.05; ** P < 0.01.

$M=1.13-0.06H_y+0.0003H_{y2}+0.05T_y-0.01T_{y2}+0.17\Delta x+0.11\Delta z$

According to Table 5, H_y, Δx, and Δz significantly affect M, while H_y2, T_y, and T_y2 show no significant impact on M. Regression analysis indicates that H_y negatively influences M, whereas Δx and Δz positively influence M. A lower H_y increases M, causing more severe stem damage. An increase in Δx and Δz also increases M.

The mechanisms of rootstock bending damaging during alignment and centring

The alignment mechanism led to complex bending damage to the rootstock. Fig. 7 illustrates the damage process when Δx and Δz were set at 6 mm. The alignment mechanism first shifted the stem from x = –6 to x = 0 along the x–axis, and then from z = –6 to z = 0 along the z–axis. The maximum bending stress region transitioned from the x to z orientation during alignment. The subsequent centring operation led to further bending of the rootstock stem. The middle portion of the stem that came into contact with

Figure 7
The development of bending damage during alignment.

the centring bar began to experience bending damage; ultimately, all sections below the centring bar suffered bending damage. The mechanical interaction indicated that the centring operation compromised the structural integrity of the middle portion of the rootstock stem, resulting in bending damage and a subsequent reduction in its load–bearing capacity. After grafting, the compromised rootstock stem would be unable to adequately support the weight of the upper scion and the grafting clip, leading to unrecoverable bending that negatively impacts grafting quality.

The effect of compression damage on grafting quality

The compression damage test are grouped into two categories: (i) deformation characteristics and compressive strain, which reveal the damage mechanism, and (ii) equivalent bending and buckling forces that demonstrate how compression damage affects grafting quality.

Deformation characteristics and compressive strain at damage position

Fig. 8 illustrates the compression deformation and subsequent rebound of 28–day–old scion stems under a 60% diameter compression ratio. Intracellular water within the stems was gradually squeezed out, accompanied by the formation of cracks during compression. As the indenter from the tensile testing machine retracted, the stems began to rebound. Meanwhile, pre–existing cracks widened gradually until complete retraction of the indenter, ultimately stabilising at a certain width. Notably, cracks consistently developed along the stem axis, oriented perpendicularly to the radial compression direction throughout the process. As shown in Fig. 9, the compressive strain of the scion stem rose linearly as the diameter compression ratio increased. The maximum strain in the scion stem exhibited an elliptical distribution in the central region, suggesting that this region is more susceptible to damage. These findings provide critical guidance for optimising the clamping and clip–feeding directions to improve grafting quality.

Figure 8
Stem deformation under different compression or retraction ratios. The retraction ratio is defined as the ratio of the indenter’s retraction distance to the initial diameter of the scion stem.

Figure 9
Compressive strain of the scion stem during the compression process.

The effects of compression damage on mechanical behaviour and success rate

The seedling age and the compression ratio significantly influenced the compression force, equivalent bending force, and equivalent buckling force (Table 6). The interaction between these two factors had a significant effect on the compression force, but not on the equivalent bending or buckling forces.

Table 6
The results of variance analysis for the compression damage test

Index	Variation source	Degree of freedom	Sum of squares	Mean square	F-value	P-value
Compression force	Seedling age	2.00	3633.01	1816.50	179.79	0.0**
	Compression ratio	9.00	12914.18	1434.91	142.02	0.00**
	Interaction	18.00	757.06	42.06	4.16	0.00**
	Model	29.00	17304.25	596.70	59.06	0.00
	Error	270.00	2727.89	10.10
Equivalent bending force	Seedling age	2.00	0.41	0.21	32.92	0.00**
	Compression ratio	9.00	0.43	0.05	7.62	0.00**
	Interaction	18.00	0.03	0.00	0.30	1.00
	Model	29.00	0.88	0.03	4.82	0.00
	Error	120.00	0.76	0.01
Equivalent buckling force	Seedling age	2.00	57.12	28.56	150.76	0.00**
	Compression ratio	9.00	5.23	0.58	3.07	0.00**
	Interaction	18.00	2.52	0.14	0.74	0.77
	Model	29.00	64.86	2.24	11.81	0.00
	Error	120.00	22.73	0.19

• Note: * P < 0.05; ** P < 0.01.

The compression force increased as the compression ratio increased (Fig. 10a), whereas the equivalent bending and buckling forces decreased (Fig. 10c and Fig. 10d). At a 40% compression ratio, the measured compression forces (10.182, 15.864, and 19.745 N for 28–, 33–, and 38–day–old seedlings, respectively) approached the respective yield limits (9.7, 15.003, and 18.154 N), as shown in Fig. 10a and Fig. 10b. Below this threshold, the compression force increased gradually, while the equivalent bending and buckling forces remained above 0.186 N and 0.91 N, respectively, significantly exceeding the scion’s maximum weight of 0.0195 N. The rate of reduction was 17% –40% for the equivalent bending force and 10% –21% for the equivalent buckling force. Under these conditions, L was 0 mm and the grafting success rate was 100%. This outcome can be attributed to the fact that at a compression ratio of less than 40%, the stem deforms elastically. As a result, its load–bearing capacity is largely preserved, allowing the grafted seedlings to maintain an upright posture and achieve high grafting quality.

Figure 10
Mechanical behaviour and the grafting success rate under different compression ratios. L is the deformation distance between the mass centre of the grafted seedling and the centre of the substrate bowl surface. GSR is the grafting success rate, which is the ratio of the number of grafted seedlings without morphological defects to the total number of seedlings. RR is the reduction rate of the equivalent bending or buckling force, which is the ratio of the difference between the pre–damage and post–damage forces to the pre–damage force

The compression force (Fig. 10a), equivalent bending force (Fig. 10c), and equivalent buckling force (Fig. 10d) all increased in older seedlings because the stems were more fibrous, leading to improved mechanical resistance. While 38–day–old seedlings exhibited the highest forces, the increase in fibrous tissue would accelerate wear on the cutting mechanism. In contrast, 33–day–old seedlings presented a good balance: they exhibited the lowest reduction in equivalent bending and buckling forces (17% and 10%, respectively) at a compression ratio of less than 40%. Considering that large–scale grafting operations often extend over a considerable timeframe, it is not practical to maintain a fixed seedling age. Therefore, a seedling age range of 33–35 days is proposed to ensure consistent graft quality.

In summary, the results presented in Fig. 10 demonstrate that high–quality grafting is ensured when the compression ratio is less than 40% and the seedling is 33–35 days old. These parameters minimise the reduction in equivalent bending and buckling forces, result in no deformation of grafted seedlings, and achieve a 100% grafting success rate.

The effect of bending damage on grafting quality

According to Table 7, seedling age significantly influenced both the equivalent bending and buckling forces, while Δx and Δz had no significant effect. This absence of an effect is attributed to the rootstock seedling being positioned inside the clamping hole (Fig. 11a and 11b) with a gap between the substrate bowl. Consequently, when the rootstock stem bent into the target position, the substrate bowl shifted its position to mitigate potential bending damage (Fig. 11c and Fig. 11d).

Table 7
The results of variance analysis for the bending damage test.

	Variation source	Degree of freedom	Sum of squares	Mean square	F-value	P-value
Equivalent bending force	Seedling age	2.00	0.21	0.11	16.93	0.00**
	Deviation distance (Δx, Δz)	2.00	0.03	0.01	2.19	0.13
	interaction	4.00	0.01	0.00	0.29	0.88
	Model	8.00	0.25	0.03	4.92	0.00
	Error	36.00	0.22	0.01
Equivalent buckling force	Seedling age	2.00	82.06	41.03	63.22	0.00**
	Deviation distance (Δx, Δz)	2.00	1.91	0.96	1.47	0.24
	interaction	4.00	0.59	0.15	0.23	0.92
	Model	8.00	84.56	10.57	16.29	0.00
	Error	36.00	23.36	0.65

• Note: * P < 0.05; ** P < 0.01.

FIGURE 11
Rootstock morphology and alignment process.

As illustrated in Fig. 12a, the equivalent bending and buckling forces decreased as the deviation distance increased. These results indicate that larger deviation distances caused more severe bending damage to the rootstock stem, thus reducing the stem’s load–bearing capacity and its ability to support the scion and graft clip, ultimately affecting graft quality and success. Specifically, when Δx and Δz exceeded 10 mm, particularly when the stem had a natural curve (Fig. 11e), the stem could deviate beyond the operational range of the alignment mechanism, leading to potential grafting failure. In contrast, when Δx and Δz were less than 10 mm, the minimum equivalent bending and buckling forces (0.064 and 0.81 N, respectively) remained well above the total weight of the scion and clip (0.0228 N). Under these conditions, the grafted seedlings exhibited no deformation (L = 0 mm), and the grafting success rate was 100%.

Figure 12
Mechanical behaviour and the grafting success rate after bending damage. GSR is the grafting success rate, which is the ratio of the number of grafted seedlings without morphological defects to the total number of seedlings. The deviation distances Δx and Δz refer to the deviation distances of seedlings in the transverse (x–axis) and longitudinal (z–axis) directions, respectively, in the upper surface of substrate bowls.

As the age of the seedling increased, the equivalent bending and buckling forces increased gradually, as shown in Fig. 12b. This trend is due to the development of thicker fibres in the rootstock stems, which enhanced bending resistance. Although selecting mature seedlings minimises mechanical damage, excessively mature seedlings can complicate cutting operations. Therefore, a moderate seedling age of 33–35 days is recommended. This age range balances structural integrity and operational feasibility, thereby promoting high–quality grafting outcomes.

Conclusions

This study investigated the effects of mechanical damage on grafting quality to optimise the operational parameters of an automatic grafting machine and seedling age. Grafting quality was assessed based on the equivalent bending force, equivalent buckling force, deformation distance, and grafting success rate. The primary sources of mechanical damage leading to graft failure were compression damage to the scion stem caused by clamping and complex bending damage to the rootstock stem induced by the alignment and centring processes. For compression damage, when the compression ratio was less than 40% and the seedling was 33–35 days old, there was minimal degradation of the stem load–capacity because the reduction rate of the equivalent bending and buckling forces reached its minimum value of 17% and 10%, respectively. For bending damage, using the finite element simulation method and the Plackett–Burman design, the main significant influencing factors were identified as the bending height and deviation distance. The optimal conditions involved the following combination: a bending height of 20 mm, a deviation distance of the rootstock emergence point in both the longitudinal and transversal directions within the substrate bowl upper surface of less than 10 mm, and a seedling age of 33–35 days. Under these conditions, the equivalent bending force (0.146 N) and the equivalent buckling force (0.978 N) were significantly higher than the weight (0.0228 N) from the graft clip and scion. Employing this combination of operation parameters and seedling age resulted in a deformation distance of 0 mm and a grafting success rate of 100%, thereby ensuring that most grafted seedlings remained upright and of high quality. This study provides a theoretical basis and data to support the design of an automatic grafting machine that minimises damage during the clamping and alignment mechanisms.

Acknowledgments

This work was supported in part by the Key-Area Research and Development Program of Guangdong Province (No.2023B0202110001).

References

• Asadbeigi, S., Ahmadi, E., Goodarzi, M., & Sagharichian, A. (2023). Analyzing and simulating heat transfer and designing a shell and tube heat exchanger for the pasteurization process of tomato paste: A CFD study. Heliyon, 9(11). http://doi.org/10.1016/j.heliyon.2023.e21593
  » http://doi.org/10.1016/j.heliyon.2023.e21593
• Chen, L., Jiang, K., Zhang, Q., Guo, W., & Zheng, W. (2020). Design and experiment on scion cutting mechanism of grafting robot for cucurbit. International Journal of Agricultural and Biological Engineering, 13 (5), 99-106. http://doi.org/10.25165/j.ijabe.20201305.5810
  » http://doi.org/10.25165/j.ijabe.20201305.5810
• Cheng, S., Wang, J., Wang, Y., & Wei, Z. (2017). Discrimination of different types damage of tomato seedling by electronic nose. (Conference).
• Comba, L., Gay, P., & Ricauda Aimonino, D. (2016). Robot ensembles for grafting herbaceous crops. Biosystems Engineering, 146, 227 - 239. http://doi.org/https://doi.org/10.1016/j.biosystemseng.2016.02.012
  » http://doi.org/https://doi.org/10.1016/j.biosystemseng.2016.02.012
• Fao (2025). FAOSTAT. https://www.fao.org/faostat/zh/#data/QCL
  » https://www.fao.org/faostat/zh/#data/QCL
• Gallegos Cedillo, V. M., Nájera, C., Signore, A., Ochoa, J., Gallegos, J., Egea Gilabert, C., Gruda, N. S., & Fernández, J. A. (2024). Analysis of global research on vegetable seedlings and transplants and their impacts on product quality. Journal of the Science of Food and Agriculture, 104 (9), 4950-4965.
• Jiang, K., Guo, W., Chen, L., Huang, W., Ge, Y., & Wei, X. (2022). Design and experiment of automatic clip-feeding mechanism for vegetable-grafting robot. Agriculture - Basel, 12 (3). http://doi.org/10.3390/agriculture12030346
  » http://doi.org/10.3390/agriculture12030346
• Jin, X., Li, D., Ma, H., Ji, J., Zhao, K., & Pang, J. (2018). Development of single row automatic transplanting device for potted vegetable seedlings. International Journal of Agricultural and Biological Engineering, 11, 67-75. http://doi.org/10.25165/j.ijabe.20181103.3969
  » http://doi.org/10.25165/j.ijabe.20181103.3969
• Kwack, Y., Park, S. W., & Chun, C. (2014). Growth and development of grafted cucumber transplants as affected by seedling ages of scions and rootstocks and light intensity during their cultivation in a closed production system. Horticultural Science and Technology, 32 (5), 600-606.
• Liang, H., Zhu, J., Ge, M., Wang, D., Liu, K., & Zhou, M., et al. (2023). A comparative analysis of the grafting efficiency of watermelon with a grafting machine. Horticulturae, 9 (5). http://doi.org/10.3390/horticulturae9050600
  » http://doi.org/10.3390/horticulturae9050600
• Liu, Z., Yang, J., Fadiji, T., Li, Z., & Jiheng, N. (2023a). Biomechanical response of the root system in tomato seedlings under wind disturbance. Phyton, 92, 1071 - 1090. http://doi.org/10.32604/phyton.2023.026408
  » http://doi.org/10.32604/phyton.2023.026408
• Liu, Z., Yang, P., Fadiji, T., Li, Z., & Ni, J. (2023b). Biomechanical response of the above-ground organs in tomato seedling at different age levels under wind-flow disturbance. Scientia Horticulturae, 312. http://doi.org/10.1016/j.scienta.2023.111835
  » http://doi.org/10.1016/j.scienta.2023.111835
• Medland, L. (2021). 'There is no time': Agri-food internal migrant workers in Morocco's tomato industry. Journal of Rural Studies, 88, 482 - 490. http://doi.org/10.1016/j.jrurstud.2021.04.015
  » http://doi.org/10.1016/j.jrurstud.2021.04.015
• Nifakos, N. (2022). Analysis of KED-Mediated wound response to biotic stress and mechanical damage in tomato plants (Solanum lycopersicum).
• Sun, Y., Wang, J., & Cheng, S. (2016). Differentiation of tomato seedlings with different mechanical damage severities using E-Nose and GC-MS. Transactions of the Asabe, 59 (5), 1069 - 1078. http://doi.org/10.13031/trans.59.11610
  » http://doi.org/10.13031/trans.59.11610
• Wu, K., Lou, J., Li, C., & Li, J. (2021). Experimental evaluation of rootstock clamping device for inclined inserted grafting of melons agriculture (11 pp.).
• Xie, L., Wang, P., Luo, J., Yi, W., & Deng, J. (2022). Optimisation and numerical simulation of shearing blade used for citrus seedling grafting. Biosystems Engineering, 215, 67 - 79. http://doi.org/10.1016/j.biosystemseng.2022.01.006
  » http://doi.org/10.1016/j.biosystemseng.2022.01.006
• Yan, G., Feng, M., Lin, W., Huang, Y., Tong, R., & Cheng, Y. (2022). Review and prospect for vegetable grafting robot and relevant key technologies agriculture (12 pp.).
• Yang, P., Hao, J., Li, Z., Tchuenbou-Magaia, F., & Ni, J. (2024). Wind disturbance-based tomato seedlings growth control. Biosystems Engineering, 243, 82 - 92. https://doi.org/10.1016/j.biosystemseng.2024.05.007
  » https://doi.org/10.1016/j.biosystemseng.2024.05.007

• Data Availability Statement:
  The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Edited by

• Area Editor:
  Gizele Ingrid Gadotti

• Edited by
  Sbea

Data availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Publication Dates

• Publication in this collection
  23 Jan 2026
• Date of issue
  2026

History

• Received
  25 Mar 2025
• Accepted
  14 Oct 2025